[GH-ISSUE #15275] Inform user of manifest results before pull #35531

Open
opened 2026-04-22 20:05:59 -05:00 by GiteaMirror · 1 comment
Owner

Originally created by @maritaria on GitHub (Apr 3, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/15275

When running a ollama pull some-model you tend to see:

> ollama pull some-model
pulling manifest 
pulling ${hash}: 0% ...

But after pulling the manifest there is some useful information to present to the user.

  1. Show the resolved final name of the model being pulled, especially when specifying an untagged model name "abc" instead of "abc:9b".
  2. Display the number of files to download and, if known, the total download size.

It could look like this:

> ollama pull some-model
pulling manifest
using some-model:9b
pulling 9 files, totaling 8.3GB
pulling ${hash}: 0% ...

Or if at the point of pulling the manifest the full model name is already resolved:

> ollama pull some-model
pulling manifest for some-model:9b
pulling 9 files, totaling 8.3GB
pulling ${hash}: 0% ...
Originally created by @maritaria on GitHub (Apr 3, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/15275 When running a `ollama pull some-model` you tend to see: ``` > ollama pull some-model pulling manifest pulling ${hash}: 0% ... ``` But after pulling the manifest there is some useful information to present to the user. 1. Show the resolved final name of the model being pulled, especially when specifying an untagged model name "abc" instead of "abc:9b". 2. Display the number of files to download and, if known, the total download size. It could look like this: ``` > ollama pull some-model pulling manifest using some-model:9b pulling 9 files, totaling 8.3GB pulling ${hash}: 0% ... ``` Or if at the point of pulling the manifest the full model name is already resolved: ``` > ollama pull some-model pulling manifest for some-model:9b pulling 9 files, totaling 8.3GB pulling ${hash}: 0% ... ```
GiteaMirror added the feature request label 2026-04-22 20:05:59 -05:00
Author
Owner

@samrusani commented on GitHub (Apr 12, 2026):

Hi! I'd be happy to work on this if a maintainer is open to it.

I've prototyped a minimal implementation — a single status line emitted in PullModel() (in server/images.go) after the manifest is resolved and layers are collected, before downloads begin:

fn(api.ProgressResponse{Status: fmt.Sprintf("pulling %s with %d files (%s total)", name, len(layers), format.HumanBytes(mf.Size()))})

Which would produce output like:

pulling manifest
pulling llama3.2:latest with 9 files (4.9 GB total)
pulling 68e0ec597aee: 100% ...

The change is 3 lines (1 import + 2 code) and uses the existing format.HumanBytes helper. Happy to adjust the wording or approach based on feedback before opening a PR.

<!-- gh-comment-id:4231951292 --> @samrusani commented on GitHub (Apr 12, 2026): Hi! I'd be happy to work on this if a maintainer is open to it. I've prototyped a minimal implementation — a single status line emitted in `PullModel()` (in `server/images.go`) after the manifest is resolved and layers are collected, before downloads begin: ```go fn(api.ProgressResponse{Status: fmt.Sprintf("pulling %s with %d files (%s total)", name, len(layers), format.HumanBytes(mf.Size()))}) ``` Which would produce output like: ``` pulling manifest pulling llama3.2:latest with 9 files (4.9 GB total) pulling 68e0ec597aee: 100% ... ``` The change is 3 lines (1 import + 2 code) and uses the existing `format.HumanBytes` helper. Happy to adjust the wording or approach based on feedback before opening a PR.
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#35531